Automatically selecting thematically representative music

Abstract

A method for automatically selecting thematically representative music is disclosed. A processor is used for using a theme-related keyword to search a keyword-indexed video repository to retrieve videos associated with the theme-related keyword; analyzing the retrieved videos to select videos with music; and extracting music tracks and features from the selected videos. The method further includes selecting representative music related to the theme from the extracted music tracks using the extracted features; and storing the selected representative music in a processor accessible memory.

The invention claimed is: 
     
       1. A method for automatically selecting thematically representative music, comprising using a processor for:
 (a) using a theme-related keyword to search a keyword-indexed video repository to retrieve videos associated with the theme-related keyword; 
 (b) analyzing the retrieved videos to select videos with music; 
 (c) extracting music tracks and features from the selected videos; 
 (d) selecting representative music related to the theme from the extracted music tracks using the extracted features; and 
 (e) storing the selected representative music in a processor accessible memory, 
 wherein the extracting features comprises extracting visual features including a cluster defining metric. 
 
     
     
       2. The method of  claim 1  where the search in element (a) uses multiple keywords, associated keywords or synonyms, boolean keyword logic or regular expressions of keywords. 
     
     
       3. The method of  claim 1  where the theme is ethnographic, geographic, demographic, or based on an event. 
     
     
       4. The method of  claim 1  further including determining the theme using analysis of images from an image collection. 
     
     
       5. The method of  claim 1  where the keyword indexed video repository comprises videos contributed from multiple sources. 
     
     
       6. The method of  claim 1  wherein the analysis in element (b) includes using a classifier trained to recognize music and non-music videos. 
     
     
       7. The method of  claim 1  wherein features in element (c) are derived from audio data, video data, or metadata. 
     
     
       8. The method of  claim 1  wherein the selection in element (d) includes audio feature-based clustering or a combination with video and metadata features. 
     
     
       9. The method of  claim 1  wherein the storing in element (e) includes storing an identifier of the selected representative music rather than the music data itself. 
     
     
       10. The method of  claim 1  further including using the selected representative music in a video or a slide show of images. 
     
     
       11. The method of  claim 1  further including using the selected representative music to replace the soundtrack in a video. 
     
     
       12. The method of  claim 1  further including purchase of thematically representative music tracks in response to thematic search requests. 
     
     
       13. The method of  claim 1  further including production of a music album of a theme based on multiple thematically representative music tracks. 
     
     
       14. The method of  claim 1  wherein the cluster defining metric comprises cluster cardinality. 
     
     
       15. The method of  claim 1  wherein the cluster defining metric comprises intra-cluster distance.